LoRaWAN, IoT & Synchronization

LoRaWAN, IoT & Synchronization ITSF 2015

Richard Lansdowne, Senior Director Network System Solutions

.

Agenda

Introduction to LoRaWAN The LoRa Alliance

Radio Parameters

Network Architecture

Classes of devices

Summary

Impacts on Synchronisation Frequency accuracy

Time-slotted access (class B)

Localisation and the requirements on synchronisation

Summary

Semtech Corporation 2 Nov 2015

LoRa-Alliance.org

LoRa Alliance™ and LoRaWAN ™

LoRa-Alliance.org

https://www.lora-alliance.org/



LoRa-Alliance.org

LoRa™Alliance • Open & Non-Profit,

Founded: March 2015

• Board

• IBM, Cisco, Bouygues Telecom, KPN, Semtech, Proximus, Actility, Kerlink,Agutek, Homerider, Sagemcom

• Chairman: Geoff Mulligan

• Committee Chairs

• Technical: Nicolas Sornin, Semtech

• Marketing: Tracy Hopkins, Stream Technologies

• Strategy: Stephen Cadwell, Microchip

• Certification: Derek Hunt, Semtech

• Specification: LoRaWAN R1.0

LoRa Alliance™: >140 members Specification downloads: >3000

LoRa-Alliance.org

LoRaWAN: 9 Announced Deployments… three live

5


True location Low power

In/out door

Low cost

Bidirectional Bidirectional

Time slotted

Broadcast

Global mobility True mobility

Seamless

Roaming

Security Full encryption

Authentication

Unique ID

LoRaWAN Network Architecture

7 Semtech Corporation Nov 2015

Sensor-Gateway (Base station) Interface

ALL GW receive on ALL channels ALL of the time

Zero network overhead- Sensor transmits a packet at any time

No GW controller necessary- Operate on same channels

Add capacity with more GW, anytime anywhere. No reuse planning reqd.

Sensors operate continuous low power, vary data rates

Super-low peak current allows 10 years on a ‘coin cell’


Network Solution

Network Server removes redundancy where required

A single packet received on many antennas can be processed if reqd.

GW are dumb & perform no data validation uplink or downlink

Network behaviour is determined by the server per application, low cost GW

Separation of data by application or user in server

Single or multiple providers per network allow choice by application


LoRaWAN device classes

Class name Intended usage

A

(« all »)

Battery powered sensors (or actuators with no latency constraint)

Most energy efficient communication class.

Must be supported by all devices

Mainly uplink with two

potential downlink slots

after each uplink

B (« beacon »)

Battery powered actuators

Energy efficient communication class for latency controlled downlink.

Based on slotted communication synchronized with a network beacon.

Programmed downlink

slots to allow control within

certain latency limits

C (« continuous »)

Mains powered actuators

Devices which can afford to listen continuously.

No latency for downlink communication.

Lowest latency command

and control for less power

critical devices.


Class A- Bidirectional communication;

END DEVICES

NETWORK

End Devices

transmit at any

time (ALOHA) Programmed wait 1

Rx

slot

1

Rx

slot

2

Time

Programmed wait 2

Receiver Initiated Transmission strategy (RIT)

Uplink Received by multiple gateways

Network Server selects: (1) Gateway for downlink (2) Which downlink slot to use


For every uplink, there are two possible downlink slots. Downlink is

possible only at these times.

Class B- Bidirectional communication;

END DEVICES

NETWORK

Pre-programmed RX slots synchronized by gateway beacons

Beaco

n

Time

Coordinated Sampled Listening (CSL)


Rx

slo

t

Rx

slo

t

Rx

slo

t

Rx

slo

t

Rx

slo

t

Rx

slo

t

Rx

slo

t

Rx

slo

t

Rx

slo

t

Be

ac

on

Rx

slo

t

Rx

slo

t

Rx

slo

t

Network Server selects: (1) Gateway for downlink (2) Which downlink slot to use

There are pre-programmed downlink slots. Downlink is possible at any of these times. Application dependent.

Summary on LoRaWAN

Highest link budget- for deep indoor coverage ‘better than GSM range’ with 25mW output power and ultra-low peak battery current

Simple deployment & densification All gateways operate on same channels- simply ‘switch on’ a new one to increase density

In-built support for localisation Time-stamping very slow signals is inaccurate, LoRa spread-spectrum allows time-stamping

Simple, low cost sensors Lowest cost electronics and battery- similar cost to Zigbee but with much greater range

National Networks Now Being rolled out Roaming, and world-wide coverage are all being addressed by The Alliance


Frequency Accuracy and Synchronisation

LoRaWAN uses ‘chirp spread-spectrum’- highly tolerant Unlike ultra-narrow-band IoT technologies, frequency offset is not a problem

Low power, low cost & low accuracy crystals are perfect for Class A & Class C

Also Mobile devices with Doppler effects are no problem

Class B requires 1us timing accuracy at each Gateway Similar requirements to many cellular applications

Local oscillator and synchronisation requirements can easily map across from cellular

Unlike cellular, better accuracy does not map to ‘more efficient’, either it works or it doesn’t

Class B also requires sensor to keep time Beacons from gateways are on a 128s period

The node must keep sub-micro-second time in between beacons


Localisation- ‘Game Changing IoT’ … driving synchronisation requirements

Reception by multiple GW enables zero overhead localisation

A well designed network can locate ANY sensor without GNSS Zero additional cost

Zero additional power

Indoors and outdoors

How?

… Time-stamping of packet arrivals… plus other parameters Combination of differential time of arrival and received signal strength plus ….


LoRa & Localisation: Problem number 1 Accurate time-base to the Gateway

Required accuracy-

‘as good as possible’

10ns error = 3 metres error!

What’s important, & what’s not?

Absolute time is NOT

Local consistency IS- differential time-stamping is used

=> GNSS receivers absolute accuracy is likely good enough

=> Other options-

PTP, ranging between GW, ensemble calculations etc.


GNSS?

PTP? Ranging?

eLoran?

LoRa & Localisation: Problem number 2 How to time-stamp accurately enough?

Required accuracy-

‘as good as possible’

10ns error = 3 metres error!

LoRa is a 150kHz spread-spectrum signal

=> Sample interval of ~ 6.7us

Semtech has developed over-sampling I.P. (256x)

=> error < 30ns (narrow-band IoT is much lower bandwidth, hence it is impossible to accurately time-stamp)


LoRa & Localisation: Problem number 3 Multi-path transmission


Multipath effects cause issues: • Longer perceived delay • Weaker signal strength

GW receiving only multipath signals cause unknown errors GW deploying antenna diversity helps a lot Differential signal strength helps Complex algorithms can be deployed to gain improvement in location accuracy.

Sagemcom at M2M Innovation World


By Kind permission of Thierry Lestable of Sagemcom, some slides from M2M World: here

http://www.slideshare.net/titidelparis/io-t-sagemcom-m2minnovationworldgeotrackv08

Calculating Position


Many groups will work on localisation with different techniques and data requirements. Here, Sagemcom list: TDOA (time difference of arrival), RSS (received signal strength), DRSS (differential RSS)

Time Difference of Arrival


GPS shown as a timing reference to all gateways! Timing accuracies with (GNSS) are expected to be significantly better than 100ns (relative)

Sagemcom: Mixing Macro and Femto cells


In this example it is shown that adding together unsynchronised femtocells with the synchronised macrocells has some benefit with increasing deployment

Sagemcom: Synchronising femtocells for greater effect


Marked improvement with synchronised femtocells. Conclusion: We need an indoor femtocell synchronisation scheme to give better than 100ns alignment!

Summary: LoRaWAN requirements on synchronization

This important technology offers the combination of multi-year

coin cell battery life, geo-location with low cost sensors

In order to make full use of indoor location, dense deployment

of gateways will be necessary

Gateways need ‘better than 100ns’ alignment…

The requirements on synchronisation just got tougher!

Thank you! Semtech Corporation 24 Nov 2015

References

LoRa Alliance Introduction

LoRa Alliance YouTube channel

Sagemcom M2M Innovation world

LoRa community Overview


https://www.youtube.com/watch?v=2Y0bMX3TVi0

https://www.youtube.com/channel/UCXYkZ2wgVIoGHLkmynVlYmQ




https://www.youtube.com/watch?v=8mfgbjMM7mY

Documents

LoRaWAN, IoT & Synchronization